FlowHijack is the first dynamics-aware backdoor attack on flow-matching VLAs that achieves high success rates with stealthy triggers while preserving benign performance and making malicious actions kinematically indistinguishable from normal ones.
Neural attention distillation: Erasing backdoor triggers from deep neural networks
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
representative citing papers
BadSNN injects backdoors into spiking neural networks by adversarially tuning LIF neuron hyperparameters and optimizing triggers, achieving higher attack success than prior data-poisoning methods while remaining robust to common defenses.
GSPure removes watermarks from 3D Gaussian Splatting by isolating watermark Gaussians via view-dependent contributions and feature clustering, cutting watermark PSNR by up to 16.34 dB with under 1 dB scene quality loss.
MIST detects Trojaned DNN updates by measuring spectral deviations in pre-activation representations against a benign fine-tuning reference, achieving high accuracy across datasets and attacks after a single update.
TIGS detects backdoor-induced attention collapse in LLMs and applies content-aware tail-risk screening plus intrinsic geometric smoothing to suppress attacks while preserving normal performance.
A complete pipeline for federated unlearning via knowledge distillation for efficient removal and a GAN-integrated classifier for visual evaluation of forgetting capacity.
citing papers explorer
-
FlowHijack: A Dynamics-Aware Backdoor Attack on Flow-Matching Vision-Language-Action Models
FlowHijack is the first dynamics-aware backdoor attack on flow-matching VLAs that achieves high success rates with stealthy triggers while preserving benign performance and making malicious actions kinematically indistinguishable from normal ones.
-
BadSNN: Backdoor Attacks on Spiking Neural Networks via Adversarial Spiking Neuron
BadSNN injects backdoors into spiking neural networks by adversarially tuning LIF neuron hyperparameters and optimizing triggers, achieving higher attack success than prior data-poisoning methods while remaining robust to common defenses.
-
Can Protective Watermarking Safeguard the Copyright of 3D Gaussian Splatting?
GSPure removes watermarks from 3D Gaussian Splatting by isolating watermark Gaussians via view-dependent contributions and feature clustering, cutting watermark PSNR by up to 16.34 dB with under 1 dB scene quality loss.
-
Detecting Trojaned DNNs via Spectral Regression Analysis
MIST detects Trojaned DNN updates by measuring spectral deviations in pre-activation representations against a benign fine-tuning reference, achieving high accuracy across datasets and attacks after a single update.
-
Defusing the Trigger: Plug-and-Play Defense for Backdoored LLMs via Tail-Risk Intrinsic Geometric Smoothing
TIGS detects backdoor-induced attention collapse in LLMs and applies content-aware tail-risk screening plus intrinsic geometric smoothing to suppress attacks while preserving normal performance.
-
Forgetting to Witness: Efficient Federated Unlearning and Its Visible Evaluation
A complete pipeline for federated unlearning via knowledge distillation for efficient removal and a GAN-integrated classifier for visual evaluation of forgetting capacity.